selfplay-gpt56-p5-27235f

SelfPlay A1 checkpoints

Run: p5nv-1e795796-g-27235f. Topic-label model: us.openai.gpt-5.6-sol, reasoning effort none.

These are LoRA adapters for the Llama-3.1-8B base model, saved during the first attacker training stage (PSRO/A1). GPT, where named, supplied topic labels; these files do not contain GPT model weights.

Saved steps in this repository: 250, 260, 270.

Training status: A1 interrupted at observed step 272 by Topic API HTTP 503. The last saved checkpoint is step 270. The full PSRO/Uniform/Naive experiment did not complete, and independent benchmarks were not run. Selection follows the largest step count, not a best-quality checkpoint criterion.

Load a selected checkpoint

Install compatible torch, transformers, peft, and accelerate packages. The base model must also be downloaded; it is not included in this repository.

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

base_id = "mlabonne/Meta-Llama-3.1-8B-Instruct-abliterated"
base_revision = "368c8ed94ce4c986e7b9ca5c159651ef753908ce"
repo_id = "zhihz0535/selfplay-gpt56-p5-27235f"
checkpoint = "checkpoints/step-270"

tokenizer = AutoTokenizer.from_pretrained(base_id, revision=base_revision)
base = AutoModelForCausalLM.from_pretrained(
    base_id, revision=base_revision, torch_dtype="auto", device_map="auto"
)
model = PeftModel.from_pretrained(base, repo_id, subfolder=checkpoint)

The checkpoints in this repository are publicly accessible.

run-info.json records the settings and export changes. manifest.json contains file sizes and SHA-256 checksums. Original adapter configurations are retained under provenance/. Adapter weights were not changed.